데미스 허사비스
Co-founder and CEO of Google DeepMind; 2024 Nobel laureate in Chemistry
174 records in all
3 free questions a day, no sign-up.
Photo: John Sears CC BY-SA 4.0 Original cropped
8 basic records of 50
Episodes are real events from public sources, summarised by us.
DeepMind combined Q-learning with deep learning to create DQN, which mastered Atari games using only raw pixels. This was their first major system that proved reinforcement learning could scale to complex tasks.
LessonCombining different technical approaches can unlock solutions for complex problems that single methods cannot solve.
DeepMind joined forces with Google partly because they believed the entire organization took the questions of AI responsibility and safety seriously. They wanted a partner that shared their 'bold and responsible' mantra for transformative technology.
LessonWhen building powerful technology, align with partners who share your ethical standards and long-term vision for safety.
Demis Hassabis founded Isomorphic Labs as a spin-off from DeepMind to focus specifically on improving human health through AI-driven drug discovery. The company was established under the Alphabet Inc. umbrella to build upon the success of AlphaFold technology.
LessonA successful core technology can be the foundation for a dedicated vertical business that solves specific high-value problems.
After folding all 200 million proteins known to science, DeepMind made the system and the resulting database freely available to the public. This move allowed researchers worldwide to use the data for drug discovery and disease understanding.
LessonSharing core breakthroughs can accelerate industry-wide progress and build immense brand authority and goodwill.
Google DeepMind developed WeatherNet, a neural network system that outperformed traditional fluid dynamics models on supercomputers. It successfully predicted hurricane paths and other chaotic weather events more quickly and accurately than previous methods.
LessonNeural networks can model chaotic real-world systems more efficiently than traditional, compute-heavy mathematical simulations.
When starting DeepMind in 2010, Demis could not raise enough money and did not pay himself for a couple of years. He contrasts this with today's environment where interns earn as much as their entire initial seed round.
LessonPersistence during periods of low capital is necessary when you are ahead of the market's understanding.
Demis discussed the legacy of polymath John von Neumann, noting how his 1950s foresight into 'grown' rather than programmed machines predicted modern AI. He connected von Neumann's work on game theory and the Manhattan Project to the current high-stakes development of AI.
LessonStudying the foresight of past masters can provide a roadmap for navigating modern technological disruptions.
Demis took challenging math and theoretical computer science courses that pushed his limits. This experience taught him how to maintain focus and work through the feeling of being stuck on technical problems.
LessonHard technical training builds the grit necessary to handle the frustrations of building a business or product.
You are seeing 28 basic records. Membership answers from all 174 of this master's records.
See membershipAI and frontier tech
These records are our compilation of public statements and sources. We are not affiliated with or endorsed by Demis Hassabis or any related company. If something should come down, tell us. Takedown request form
